Re: [Axiom-developer] Re: doyen

[Alejandro Jakubi]

CY

> That's not possible in general - if nothing else, the hardware platforms on
> which software is run will not remain static.  A result turned out on a PDP-8

Certainly, bitwise reproducibility does not seem realistic.

But for the purposes of scientific documentation, a weaker form, the
reproducibility of results or black box reproducibility say (same input, same
output) is enough, but required.

The old system works that way. Undoubtly, the minds of the author(s) of a
paper/book worked differently than mine. However, the nature of the human mind
and the academic training allows me to reproduce the hand calculations 
he/she/they
made in the past. And eventually identify wrong path of thinking in case of
errors.

By the way, correcting exams is a good training in realizing wrong ways of
making calculations!

Could it be that this reproducibility of results requires, besides a suite of
formal proofs, some sort of artificial inteligence?

Anyway, my point is: if papers will be archived with its code, but after some
years that code will not run, or give results diferent to those stated in the
paper, it will not make much sense to archive papers with code for long.

To the contrary, the success of initiatives like ArXiV for archiving text and
figures is based on the stability of TeX and PostScript.


Regards, Alejandro